PUBLIC HEALTH ASSESSMENT
RIO ABAJO, HUMACAO COUNTY, PUERTO RICO
In conducting an ATSDR public health assessment, the health assessors identify and review allavailable environmental contamination data for a site. The on- and off-site portions of thissection describe the sampling that has been done and identify contaminants of concern.
The extensive environmental sampling conducted in the Frontera Creek area from 1978 - 1985 issummarized in Appendix E. The most complete, accurate, and valid sampling was that done inthe remedial investigation (RI) conducted in 1987 - 1988. Valid air monitoring was done in1993-94 by a contractor for Squibb with oversight by EPA.
The field sampling program for the RI included the collection and analysis of over 1,000 samplesof soil, sediment, surface water, groundwater, potable water, biota, and air. All samples wereanalyzed for mercury and 20 percent were checked for the compounds on EPA's HazardousSubstances List (HSL). The HSL includes over 100 organic chemicals including lindane, andalso metals. Background samples from locations in the area were collected for each media. These background locations were selected to be as similar to the site as possible.
The quality of the environmental data is discussed in the Quality Assurance and Quality Controlsection (page 18). Physical and other hazards not related to toxic substances, if any, aredescribed in the Physical and Other Hazards section (page 18). This introductory portiondiscusses the process for selecting contaminants of concern, Toxic Chemical Release Inventory(TRI) data, and the relevance of the environmental sampling data to public health.
Selection of Contaminants of Concern
ATSDR selects contaminants for further evaluation based upon the following factors:
- comparison of concentrations of contaminants on- and off-site with values for noncarcinogenic and carcinogenic endpoints,
- sampling plan and field and laboratory data quality, and
- community health concerns.
Identification of a contaminant of concern in the On-Site and Off-Site Contamination sectionsdoes not mean that exposure will result in adverse health effects, only that additional evaluationis necessary. The public health significance, if any, of exposure to the contaminants of concern isevaluated in subsequent sections of the public health assessment. If a specific contaminant isclearly not site-related, that will be mentioned. For the protection of public health, the possiblehealth impact of those non-site related contaminants will be evaluated. Once a contaminant hasbeen identified in one environmental media (e.g., ground water), the contaminant's concentrationin other environmental media is evaluated for possible combined exposures above a comparisonvalue. If those are identified, that will be reported and evaluated further.
The comparison values used in a public health assessment are contaminant concentrations inspecific media that are used to select contaminants for further evaluation. Those values includeEnvironmental Media Evaluation Guides (EMEGs), Cancer Risk Evaluation Guides (CREGs),and other relevant guidelines which are described on page 54 in Appendix A.
Typically, a public health assessment uses soil comparison values for noncarcinogenic effects forchildren. This results in the comparison values being considerably lower than if adult valueswere used. This is because a child's weight is less than an adult's and because it is assumed that achild ingests 2 - 50 times as much soil as an adult. In this public health assessment, soilcomparison values for adults will be used for evaluating the on-site data. It is unlikely that anysignificant exposure to children has occurred because of the inaccessibility of the on-sitefacilities. Comparison values for children will be used for off-site data.
Review of Toxic Chemical Release Inventory (TRI) Data
To identify facilities that may contribute to the groundwater, surface water, and air contaminationnear the Frontera Creek NPL site, ATSDR searched the 1987-1992 files of the Toxic ChemicalRelease Inventory (TRI) databases for the Humacao area (13). TRI was developed by EPA fromchemical release (air, water, and soil) information provided by certain industries.
Several limitations of TRI data should be noted (14). The air release data in TRI may beestimates or actual measurements. Most reported data are estimates based on conservative(overestimated) scenarios. Consequently, the levels of emissions recorded in TRI are oftenbiased on the high side. Another limitation of the TRI database is that only certain industriesemploying more than a minimum number of workers are required to report releases. In addition,reporting is restricted to specific chemicals that are used or released above specified amounts. Finally, it is believed that there have been and still are industries that do not report releases. Smaller industries may not be aware that reporting requirements exist or that they are responsiblefor such reports.
The search of TRI indicates that there are 7 or 8 facilities in Humacao, depending on the year,that report toxic chemical releases (13). The reported releases to the air ranged from 395,858pounds (lbs) in 1992 to 677,079 lbs in 1989. For 1987-92, one facility (Squibb Manufacturing)accounted for 69-82% (287,562 - 557,516 lbs) of the reported air releases in Humacao. For1987-92, the vast majority of Squibb's releases was methylene chloride and acetone. Thereleases of acetone varied from 109,800 lbs in 1987 to 148,110 lbs in 1989, and for methylenechloride varied from 103,300 lbs in 1987 to 233,520 lbs in 1989. Those TRI data will bediscussed further on page 21 in the Pathways Analyses and on page 39 in Community HealthConcerns Evaluation sections.
Relevance of Environmental Sampling
The sampling data described in the Environmental Contaminants and Other Hazards section arewell-suited for evaluating the possible health impact of the site. The systematic sampling done inthe remedial investigation (RI) provides good information on all the possible site-relatedenvironmental exposure pathways, except the ambient air. Additional air monitoring done in1993-94 provided adequate data, so there is good information on all the possible site-relatedenvironmental exposure pathways.
The sampling of the surface and subsurface soil and ground water in Ciudad Cristiana waswell-designed to identify exposures of public health concern. Seventy-five of the surface soilsamples were from randomly selected lots and another 72 were from lots requested by EPA,EQB, and a Citizens Advisory Board (4). For those 72 lots, an emphasis was placed on locationswhere elevated levels had been identified in EQB's 1985 sampling or where individuals withelevated mercury levels in body tissues lived. Six of the 12 locations where soil boring was doneto obtain subsurface samples, were picked because they had the six highest mercury levels in anearlier investigation (4). Boring locations were also weighted towards that part of the lot wherethe sewer or water mains were due a concern about contamination due to broken sewer and waterlines. The 12 borings were also used to obtain samples of ground water. This focus on worstcontaminant levels in the soil, subsurface soil, and groundwater sampling of Ciudad Cristianawould probably overestimate any community-wide health hazards.
Locations for the sampling of industrial facilities were selected using information fromwalk-through surveys which included use of an Organic Vapor Analyzer (OVA) and a JeromeInstrument Mercury Analyzer to identify organic chemical and mercury contamination (4). Industrial soil samples were collected from three different depths (0-6 inches, 6-18 inches, and18-36 inches). The design of the industrial soil sampling insures that the major contaminantsources were identified.
The sediment sampling program was designed to identify the lateral and vertical distribution ofcontaminants within the Frontera Creek area (4). Samples were obtained from 150 locations inFrontera Creek, the Technicon and Squibb ditches, the three Frontera lagoons, and the MandriCanal. Except for the Technicon ditch where a transect study (i.e., samples were obtained atpredetermined intervals) was done, those sample locations were depositional areas (places wheresediment settles out). At all locations, samples were taken from a depth of 0-12 inches and atabout ½ of the locations, samples were taken at a depth of 12-24 inches. The design of thesediment sampling is weighted towards finding higher concentrations.
The design of the biota sampling program focused on the animals and plants used as food ormost apt to accumulate contaminants, especially mercury (4). The selection of the biota to besampled was made through site visits, bird censuses, and bank surveys of fishermen. Differentlevels of the food chain were sampled along with domestic cattle and coconuts. Samples of thesame animals and plants were taken from two control locations whose ecosystems were verysimilar to the Frontera Creek area. Those locations were the Boqueron Wildlife Refuge, which ison the southwest corner of Puerto Rico, and Roosevelt Roads Naval Air Station, which isnortheast of Humacao.
The 1989 air monitoring program was designed to identify possible airborne contaminants, not toquantify exposure levels. A total of 11 rounds of eight-hour sampling for mercury and five forvolatile organic compounds (VOCs) were conducted from July 10 through August 18, 1989 (4). Five to seven locations were sampled for mercury and five for VOCs in each round.
Except for stormy periods, daily wind direction and speed patterns and average temperature arefairly similar year-around (4). In Humacao, the wind blows four to eight miles per hour (mph)from the east-southeast during the day and is under 2.5 mph from the north-northwest during thenight. Therefore, air sampling done anytime of the year would probably be fairly representativeof year-around exposures as long as there were an adequate number of sampling events and themonitoring accounted for the day-night changes in wind direction and speed. Unfortunately, the1989 monitoring was conducted only during the day.
Since Ciudad Cristiana is to the southeast of the industrial facilities, the air monitoring programdoes not appear to have assessed possible exposure levels at Ciudad Cristiana when they mighthave been the highest - at night. This is because, as described, the land breezes (from shore towater) at night are usually lighter than sea breezes (from water to shore) during the day, resultingin the depositing of contaminants closer to the point of release.
A bimonthly monitoring program was conducted of the Frontera Creek area by AppliedGeotechnical and Environmental Services Corporation (AGES) for Squibb Manufacturing fromApril 1993 - February 1994 (15). Sampling for methylene chloride was conducted over athree-day period every other month for a year. At least one evening and one night 8-hour spanwere sampled during each period. Six locations were sampled including one at the northern endof Ciudad Cristiana. The design of this monitoring program addresses ATSDR's objections tothe earlier monitoring.
The environmental sampling done in the remedial investigation (RI) is especially useful forevaluating human exposures to mercury in the environment. Mercury persists in the environmentfor long periods, and accumulates in the food chain with the predators at the top of the chainhaving the highest concentrations (16). Therefore, evidence of contamination should be found insoil, sediment, water, and biota. The major pathways through which humans could be exposed tomercury are eating contaminated fish and crabs, ingesting contaminated soil, or drinkingcontaminated water. The sampling of the soil, sediment, water, and biota was extensive enoughto properly identify the amount and extent of mercury contamination in the Frontera Creek area.
For this public health assessment, on-site means the industrial park and other portions of theNPL site north of Highway 3 (Figure 2, Appendix F). This section identifies those contaminantsfrom the industrial park area that meet ATSDR's guidelines for a contaminant of concern. Mercury levels above background were found only at one of the industrial sites, Technicon, andsediment from that portion of Frontera Creek immediately adjacent to Technicon. Methylenechloride and methyl isobutyl ketone were identified as contaminants of concern in on-site surfacewater. Methylene chloride, acetone, and mercury were identified as contaminants of concern inon-site air. Except for arsenic at two industrial sampling locations, levels of all other chemicalsanalyzed for at all other locations and media were below method detection limits, or were withinwhat was determined to be background for the area. However, the levels of barium, beryllium,and manganese in some areas, though within background, did meet ATSDR's guidelines for acontaminant of concern.
A total of 173 soil samples were taken from 13 industrial facilities. Only total mercury, arsenic,and beryllium were identified as contaminants of concern (Table 1, Appendix A). Arsenicconcentrations exceeded the comparison value at both sampling locations on the Coloron facility,1 of 9 locations on the Industrial Park Waste Water Treatment Plant, both locations on PeerlessTube (including the highest arsenic concentration of 430 ppb), 1 of 4 on Reedco, 2 of 5 atSquibb, and 1 of 2 on the WJK facility. Only two of the nine samples above the arseniccomparison value were greater than what is considered background (4). The single berylliumconcentration above the comparison value was on the Owens-Illinois facility.
Mercury levels exceeded the comparison value only on the Technicon facility. The highestmercury concentration at that facility was 535 parts per million (ppm). This sample wascollected immediately adjacent to a small break in a concrete berm surrounding the former rawmaterials storage area. In nearly all locations, mercury concentrations decreased substantially thedeeper in the soil that the sample was taken.
Arsenic and beryllium were identified as contaminants of concern in on-site sediment (Table 2,Appendix A). Thirty-four of the 158 on-site sediment samples, analyzed for mercury, wereabove the method detection limit of 0.08 ppm. The highest mercury concentration in sedimentwas found in the Technicon ditch. The highest mercury level in Frontera Creek sediment, 2.9ppm, was identified just downstream from the Technicon ditch.
No total or inorganic mercury was detected in groundwater samples taken from 6 monitoringwells on the industrial sites. No other analytes were evaluated.
Eight surface water samples were collected from the Technicon ditch and that portion of FronteraCreek that passes through the industrial facilities. Mercury was detected in 4/8 samples, with ahigh of 0.9 parts per billion (ppb). The EPA drinking water standard for mercury is 2 ppb.Several organic compounds were detected in 8 on-site water samples, but only methylenechloride and methyl isobutyl ketone were identified as contaminants of concern (Table 3,Appendix A).
Samples for mercury in the air were obtained from seven on-site locations using the methodologydescribed on page 12. The average level of mercury in air was 0.4 micrograms/m3 (µg/m3) with ahigh of 2.2 µg/m3 while at the background locations air levels averaged 0.04 µg/m3. Mercury inon-site air is considered a contaminant of concern.
Samples from four locations were analyzed for a variety of volatile organic compounds. Maximum concentrations of 618 µg/m3 acetone and 1956 µg/m3 methylene chloride weredetected (Table 4, Appendix A). Methylene chloride and acetone are considered contaminants ofconcern. Acetone is included even though the maximum contaminant level detected was belowthe comparison value. This is because the comparison value is for acute or short exposures andthe exposures here probably were longer, so further evaluation is needed.
Additional air monitoring for methylene chloride was conducted in July and August 1992, andbimonthly from April 1993 through February 1994 by AGES for Squibb (Table 4, Appendix A)(17,18). In 1992, three onsite locations were sampled over several 24-hour periods. Two of thethree locations were used in the 1989 remedial investigation and the third is close to a 1989location. The highest averages onsite varied from 244 µg/m3 in the evening to 498 µg/m3 nightto 1278 µg/m3 during the day with a 24-hour average of 642 µg/m3. During one of the daytimeperiods, there was a maximum methylene chloride level of 4969 µg/m3 (18).
In the bimonthly monitoring in 1993-94, four locations were sampled with the highest averagesat a location varying from 597 µg/m3 at night to 815 µg/m3 evening to 1027 µg/m3 daytime (17). Concentrations for an eight-hour period varied from nondetect - 3,430 µg/m3 daytime, nondetect- 2,600 µg/m3 evening, and 40 - 2,200 µg/m3 night
In this public health assessment, off-site is the Ciudad Cristiana subdivision plus those areasoutside the Frontera Creek NPL site boundaries. Ciudad Cristiana is actually part of the NPLsite. This section identifies those contaminants from the off-site area that meet ATSDR'sguidelines for a contaminant of concern. Mercury levels in surface soil were above thecomparison value at only one sampling location at Ciudad Cristiana. Levels of arsenic, barium,beryllium, and manganese in Ciudad Cristiana surface soils exceeded comparison values at mostsampling locations. Mercury and methylene chloride were identified as contaminants of concernin off-site air. Levels of all other chemicals analyzed for all other media were below methoddetection limits, or did not meet ATSDR's guidelines for a contaminant of concern.
A total of 164 surface and 79 subsurface soil samples were collected in Ciudad Cristiana andanalyzed for total mercury. Seventy-five of the surface soil samples were from randomlyselected lots and the rest were from lots requested by EPA, EQB, and a Citizens Advisory Board. Included in this last group were samples from the lots identified in the February and March 1985surveys done by EQB as having the highest levels of mercury. Thirty-two of the 164 surface soilsamples were analyzed for the substances on EPA's Hazardous Substance List (HSL). Ten of the79 subsurface soil samples were tested for the HSL compounds.
In Ciudad Cristiana, mercury concentrations in surface soil (0-3 inches) samples ranged from 0 to0.836 ppm with an average of 0.096 ppm. Distribution was as follows: 73 of 164 (44.5%)surface soil samples were below the method detection limit of 0.08 parts per million (ppm), 83 of164 (50.6%) between 0.08 to 0.2 ppm, 6 of 164 (3.7%) between 0.2 to 0.3 ppm, and 2 of 164(1.2%) above 0.3 ppm. The two values above 0.3 ppm were 0.332 and 0.836 ppm. The 35background surface soil samples ranged from below the method detection limit to 0.15 ppm withan average mercury level of 0.057 ppm. Soil mercury concentrations detected at CiudadCristiana were all within the range of values reported to occur naturally in soils (4).
One of the 164 surface soil samples from Ciudad Cristiana had a mercury concentration abovethe soil mercury comparison value, which makes mercury a contaminant of concern. Additionalcontaminants of concern in Ciudad Cristiana surface soil were arsenic, barium, beryllium, andmanganese (Table 5, Appendix A).
Mercury levels in 6 of the 79 samples of subsurface soils (greater than 3 inches deep) fromCiudad Cristiana were above the method detection level of 0.08 ppm with a high of 0.236 ppmand an average of 0.023 ppm. In contrast, 3 of the 40 background samples were above themethod detection level with a high of 0.109 ppm and an average of 0.037 ppm. Thecontaminants of concern in subsurface soil from Ciudad Cristiana were barium, beryllium, andmanganese (Table 6, Appendix A).
The presence of dredge spoils (i.e., creek sediments) in the underlying fill was carefully checkedfor during the boring of the ground water monitoring wells in Ciudad Cristiana (4). No dredgespoils/sediments were identified.
Twenty-three of the 100 off-site sediment samples, analyzed for mercury, were above the methoddetection limit of 0.08 ppm. The highest concentration of the samples from the off-site portionof Frontera Creek was 2.0 ppm, from the Frontera Lagoons it was 0.33 ppm, and from theMandri Canal all samples were below the method detection limit.
Samples of ground water were taken from 12 monitoring wells in Ciudad Cristiana. Three ofthose 12 monitoring wells had detectable levels of mercury ranging from 0.22 to 0.33 ppb.
Two potable water samples were collected from a Cristiana hydrant and tested for mercury andHSL parameters. As described in the Background section, Ciudad Cristiana received water fromthe Humacao Municipal Water Supply which obtains most of its water from the Humacao River. No mercury was detected. Chloroform, at about 68 ppb, was the only HSL parameter thatexceeded its comparison value of 6 ppb. The chloroform concentration did not exceed the EPAdrinking water standard for chloroform of 100 ppb.
Mercury was not detected in the 10 samples of surface water collected from Frontera Creek andthe 4 from the Frontera Lagoons. In the 14 samples analyzed for HSL compounds, detectablelevels of acetone, carbon disulfide, ethylbenzene, methyl ethyl ketone, methylene chloride,methyl-iso-butyl ketone, xylene, toluene, benzoic acid, barium, chromium, lead, and zinc werefound. None of those exceeded their comparison values.
As described in the introductory portion of this section, mercury concentrations in air weredetermined by using a low flow pump to pass air through a mercury dosimeter. The dosimeterwas then analyzed with a Jerome Instrument Mercury Analyzer. This semiquantitative methodwas employed because there was no specific EPA method for mercury in air. Samples weretaken from two locations in Ciudad Cristiana and two other off-site locations. Sampling appearsto have been conducted only during the day.
The average levels of mercury in air from the four off-site locations ranged from 0.043 and 0.087µg/m3 (Table 7, Appendix A). Concentrations at the background locations averaged 0.04 µg/m3. The EPA has established 1 µg/m3 of mercury as the National Emission Standard for HazardousAir Pollutants (NESHAPS) level. Acetone was detected at one of the locations. Methylenechloride was detected at all four off-site locations. Mercury and methylene chloride are bothconsidered contaminants of concern (Table 7, Appendix A). Acetone will also be evaluatedfurther in this public health assessment because it is known to enhance the toxic effects ofmethylene chloride. In addition, acetone will be evaluated further because the air levels atCiudad Cristiana could have been much higher based on the wind blowing from the source toCiudad Cristiana at night.
As discussed in the Environmental Contamination and Other Hazards section (page 12), airsampling was not conducted at night during the 1989 remedial investigation (4). However,additional monitoring for methylene chloride was conducted in 1992 and 1993-4 includingsampling at night (17). In 1992, the average levels from one location varied from 2.8 µg/m3during the day to 3.9 µg/m3 evening to 29 µg/m3 at night, with a 24-hour average of 12 µg/m3. Two locations were sampled in 1993-4. At Ciudad Cristiana, which is to the southeast of thesite, the day and evening averages were 4.2 µg/m3, the night average was 8.5 µg/m3, and the24-hour average 5.6 µg/m3. Concentrations for an eight-hour period varied from nondetect - 9µg/m3 daytime, nondetect - 14 µg/m3 evening, and nondetect - 21 µg/m3 night. At the otheroffsite location which is to the southwest, average levels ranged from 2.1 µg/m3 in the day to 25 µg/m3 evening to 60 µg/m3 at night, while 24-hour average was 29 µg/m3. Concentrations foran eight-hour period varied from nondetect - 8 µg/m3 daytime, nondetect - 84 µg/m3 evening, andnondetect - 143 µg/m3 night.
The selection of the biota to be sampled was made through site visits, bird censuses, and banksurveys of fishermen (4). Different levels of the food chain were sampled along with domesticcattle and coconuts. The major biota sampled were the common gallinule (a game bird), cattleegret, tilapia (an omnivorous food fish), tarpon (a top-predator fish), shrimp, and blue crabs. Sample collection was done from the habitats nearest the site where the specific species resided.
Only 3 of the 22 bird samples had mercury levels above the method detection limit with a high of132 ppb. Six of 34 tilapia had levels above the method detection limit. However, all 6 of thesewere fish from Frontera Creek with concentrations ranging from 64 to 460 ppb. Seventeen of 29tarpon had mercury levels above the method detection limit with a high value of 244 ppb. However, the fish from the background areas had somewhat higher levels than those fromFrontera Creek. Mercury levels in the crab or shrimp samples were below the method detectionlimit.
The laboratory quality control/quality assurance protocol prescribed by EPA Region II wasfollowed including the collection of replicate samples (4). There was a 100 percent split ofsurface soil samples from Ciudad Cristiana between the responsible party's contractor (Dynamac)and the contractor for EPA (NUS). In addition, there was a 25 percent split of surface soilsamples between Dynamac and the contractor for EQB (IT Corporation). A statistical evaluationof the laboratory analyses of the split samples by three independent laboratories identified nosignificant difference. This result strongly indicates that the laboratory analysis of the CiudadCristiana surface soils was properly and accurately done.
Hazards which may exist at this site are typical of hazards that would normally exist at operatingindustrial parks. The abandoned subdivision may contain some hazards. However, there isdrive-by patrolling of the area which should aid in limiting the number of trespassers.
In this section of the public health assessment, the possible exposure pathways are evaluated tohelp determine whether persons have, are, or will be exposed to contaminants associated with thesite. Pathway analysis consists of five elements (2):
- identifying contaminants of concern possibly related to the site,
- determining that contaminants have/are/will be transported through an environmental medium,
- identifying a point of exposure (i.e., is there a place or situation where humans might beexposed to the contaminated media?),
- determining that there is a plausible route of human exposure (i.e., can the contaminant enter the body?), and
- identifying an exposed population (i.e., how many people, if any, are at the point ofexposure?).
An exposure pathway is considered complete when there is good evidence that all five elementsexist. The presence of a completed pathway indicates that human exposure to contaminants hasoccurred, is occurring, or will occur. When one or more of the five elements of an exposurepathway are missing, the pathway is considered potential. The presence of a potential exposurepathway indicates that human exposure to contaminants could have occurred, could be occurring,or could occur. An exposure pathway can be eliminated if at least one of the five elements isabsence and will never be present. If there is uncertainty about the site-relatedness of thecontaminants of concern in an exposure pathway, the pathway will be evaluated as if it weresite-related. The completed exposure pathways are presented in Table 8 (page 62), potentialexposure pathways in Table 9 (page 63), and estimates of the number of exposed individuals forthe Frontera Creek site in Table 10 (page 64).
There are four completed surface soil exposure pathways - mercury on the Technicon facility,arsenic on 6 industrial facilities, beryllium on the Owens-Illinois facility, and 5 metals on CiudadCristiana. Each pathway is considered completed because there is a strong likelihood thathumans have contacted contaminants of concern in surface soil. The exposure pathway atTechnicon facility is considered site-related, while the Ciudad Cristiana, Owens-Illinois, and theindustrial facility pathways are not. Each pathway will be discussed in more detail.
Technicon Soil Exposure Pathway
The Technicon facility has a completed soil exposure pathway that involves a single contaminantof concern - mercury. The sampling results indicate that the contamination is widespreadthroughout the facility. However, there is only one location at which the total mercury levelexceeded the comparison value for adult exposures to mercury. This location is immediatelyadjacent to a small break in a concrete berm surrounding the raw materials storage area wheremercury-containing compounds and other materials were stored (4).
This location and other contaminated areas at Technicon have been covered with clean soil, sohuman exposure should no longer occur (3,10). When those areas were not covered, Techniconemployers working around the raw materials storage area would be at risk of significant exposureif they ingested or had dermal contact with the contaminated soil. It is very unlikely that childrenor individuals not working for Technicon could have been exposed because the Techniconfacility is fenced and guarded. The maximum exposure length to mercury in this pathway wouldhave been about twenty years. The Technicon facility opened in 1971 and access to thecontaminated areas stopped in 1990 (3,4).
The possible health impact of this exposure pathway are evaluated on page 27 in the PublicHealth Implications section.
Industrial Soil Exposure Pathway
There are 6 facilities in the industrial area - Coloron, Humacao Industrial Park Waste TreatmentPlant, Peerless Tube, Reedco, Squibb, and WJK, where there is a completed soil exposurepathway for arsenic. Only Coloron (13 ppm) and Peerless Tube (430 ppm) had samples (oneeach) which were above the U.S. background level for arsenic. Based on this, there does notappear to be widespread contamination.
Exposure to arsenic at those 6 locations is still occurring. Exposure is probably restricted toworkers at the facilities. The waste treatment plant, Squibb, and Reedco are known to be fenced. The maximum exposure length to arsenic in this pathway is about 20 years. Most of the facilitiesopened in the early 1970's.
The possible health impact of this exposure pathway are evaluated on page 25 in the PublicHealth Implications section.
Owens-Illinois Facility Exposure Pathway
There is a completed exposure pathway for beryllium at the Owens-Illinois facility. Theberyllium concentration (0.3 ppm) at 1 of 2 sampling locations at this facility was above
the beryllium comparison value, but is within the background range for this metal.
Exposure to beryllium at this one location is still occurring. Exposure is probably restricted toworkers at the Owens-Illinois facility. It is not known whether this facility is fenced. Themaximum exposure length to beryllium in this pathway is about 20 years.
Ciudad Cristiana Soil Exposure Pathway
There are five metals in the Ciudad Cristiana soil exposure pathway - arsenic, barium, beryllium,manganese, and mercury. Mercury concentrations above the comparison value were found at 1of 164 locations. In contrast, nearly all of the samples for the other four metals were above thecomparison values for each chemical. However, all the concentrations of arsenic, barium,beryllium, and manganese were within the range considered background for those metals. Therefore, the source of those metals is probably natural.
Exposure to the five metals in Ciudad Cristiana soil occurred for a maximum of 6½ years, 1979through May 1985. Based on data collected from many of the Ciudad Cristiana households bythe former residents group, the average residency was 4.3 years. (The collection of thisinformation on residency is described on page 32 in the Health Outcome Data Evaluationsection.) All residents of Ciudad Cristiana likely had contact with soil with levels of arsenic,barium, beryllium, and manganese above the comparison values for each metal. However, it isprobable that only a few residents made contact with soil from that one location where themercury level was above the mercury comparison value. This sampling location was a testboring site next to a house on the very southeast corner of Ciudad Cristiana (the farthest point inthe subdivision from the industrial facilities and from Frontera Creek (Figure 2). The samplewas from the first 3 inches of soil from the boring of a groundwater monitoring well. Acomposited surface soil sample, obtained from the backyard of this same house, had a mercurylevel of 0.112 ppm. Samples from elsewhere in this corner of Ciudad Cristiana were all below0.4 ppm.
The possible health impact of this exposure pathway are evaluated pages 26 - 28 in the PublicHealth Implications section.
Ambient Air Pathway
There is an ambient air human exposure pathway in the Frontera Creek area based on the resultsof the air monitoring data done. The point of exposure is downwind of the point of release. Based on the TRI data described in the Environmental Contaminants and Other Hazards section,the Squibb facility is the main point of release for acetone and methylene chloride. It is unclearas to the source of the airborne mercury.
With Squibb as the principal source, the major exposure point to acetone and methylene chlorideat night would have been Ciudad Cristiana. This is because at night the wind blows to thesouth-southeast (see Figure 2). Since at night the wind speed is low (under 2.5 mph), themajority of the contaminants would be deposited close to the point of release. Sampling donein 1992 and 1993-4 confirms that night time levels were higher. The night time average level ofmethylene chloride in 1992 at the northern edge of Ciudad Cristiana was 10 times greater thanduring the day (29 µg/m3 vs. 2.8 µg/m3 ). A similar difference was observed in the 1993-4monitoring, but levels were lower due to implementation of a emission reduction program (17).
There are no monitoring data for the period (1979 - 1985) when people lived in Ciudad Cristiana,so indirect evidence must be used to estimate possible exposure levels. A report by Squibb'senvironmental consultant, Applied Geotechnical and Environmental Service Corporation(AGES) indicates that methylene chloride usage was limited to one manufacturing process until1983 (18). Therefore, methylene chloride levels during 1979 - 1982 were probably significantlylower than the levels measured in 1992.
However, TRI data suggest that methylene chloride levels in the air during 1983 - 1985 couldhave approached the levels observed in 1992 (13). The total amount of methylene chloridereported released in 1987 at the Squibb facility was 103,300 pounds which is similar to thereported release in 1992 of 118,800 pounds. The amount of non-point source releases was aboutsix times greater in 1987 (65,500 lbs.) than in 1992 (11,600 lbs.). The portion of total airreleases that were due to non-point sources were over 60% for 1987-89 and about 50% for 1990,while they were about 10% for 1991-92. Non-point source releases would have the greatestimpact on those areas such as Ciudad Cristiana which are immediately downwind of the source. Therefore, the airborne concentrations of methylene chloride in 1983 - 1985 could have easilyapproached the levels observed in 1992, if we assume that the amount of methylene chloridereleases in 1983 - 1985 were similar to the releases in 1987 and 1992. Exposure would havebeen for a maximum of 6½ years (1979 - May 1985) with air levels of methylene chlorideprobably significantly lower in 1979 - 1982 than 1983 - May 1985.
During the day, the wind blows to the west-northwest at somewhat higher speeds than at night(4-8 mph). Because the wind speed is higher, the disposition of the contaminants would be overa wider area. Therefore, the point of exposure during the day would be the industrial facility areaand the residential area to the west - Rio Abajo. It is assumed that exposure occurred nearlyeveryday from the probable initiation of use of methylene chloride at Squibb in 1975 through thepresent time. The maximum levels observed on-site in 1989 were 2.2 µg/m3 for mercury, 618µg/m3 for acetone, and 1956 µg/m3 for methylene chloride. The highest onsite methylenechloride levels were 4989 µg/m3 in 1992 and 3430 µg/m3 in 1993-4. Both those elevated levelswere attributed to an air stagnation zone next to the Squibb production buildings (19). Thisindicates that exposures significantly greater than the average levels could have occurred.
Methylene chloride exposure levels for Rio Abajo can be estimated from the results for the onsitesampling location on the western edge of the site. Day time and 24-hour average levels were 85and 46 µg/m3 in 1992 and 106 and 60 µg/m3 in 1993-4, respectively. As discussed previously,air levels in 1983 - 1985 likely were similar to the levels identified in 1992. Air concentrationsbefore 1983 were probably significantly lower than those in 1992.
As reviewed on pages 14 and 16 - 17, the existing air monitoring data indicate that there is greatvariation in methylene chloride concentrations. Therefore, exposure to higher levels would beepisodic. For example, during one 24-hour sampling period in April 1993, eight-hour averagemethylene chloride levels at one on-site location were 2000, 2200 and 2600 µg/m3 (17). Incontrast, in December 1993 the eight hour averages were 40, 60, and 60 µg/m3.
Surface Water Pathway
Surface water represents a completed exposure pathway for methylene chloride and methylisobutyl ketone based on environmental data. The point of exposure is that portion of FronteraCreek that flows through the industrial area. Workers and other individuals frequenting theindustrial area could have been exposed by drinking or having skin contact with thecontaminated water in the on-site portion of Frontera Creek.
Potential exposure pathways indicate that exposure to a contaminant is, was or could be possible,but is not definite.
On-site sediment represents a potential exposure pathway for arsenic and beryllium because fourof the five elements that form a completed pathway exist. The missing element is informationidentifying who has, is, or will be exposed; and the circumstances of exposure (i.e., daily lengthand level of exposure). The point of exposure is that portion of Frontera Creek that runs throughthe industrial area. Workers and other individuals frequenting the industrial area could have beenexposed by ingesting or having skin contact with contaminated sediment from the on-site portionof Frontera Creek. For this exposure pathway to be of health significance, individuals wouldhave to ingest or have skin contact with the creek sediments on a daily basis. Because there is noindication of this occurring, this pathway will not be evaluated further.
Eliminated exposure pathways are indicated if exposure to a site-related contaminant neveroccurred, or will not occur.
Ground water represents an eliminated exposure pathway because there is no indication that theground water from the Frontera Creek area is used as a drinking water source. In addition, the available data indicate that there is no significant contamination of the FronteraCreek ground water by mercury or HSL chemicals. This pathway will not be evaluated further.
Food Chain Pathway
The food chain represents an eliminated exposure pathway because there is no indication ofsite-related contamination of the food chain (i.e., area animals and plants that are eaten). Mercury was identified in area fish, but at levels similar to other areas of Puerto Rico. Mercurycontamination of area fish will be discussed further on pages 74 - 75 of Appendix D because,while the mercury is not site-related, the ingestion of contaminated fish could be a health hazard.
Potable water (Humacao Municipal Water Supply) represents an eliminated exposure pathwaybecause there is no indication of contamination of this water supply by site-related contaminants. The Humacao water supply obtains most of its water from the Humacao River which has noconnection with the Frontera Creek area drainage basin. The chemicals identified in thesampling of the potable water did not exceed EPA's drinking water standards. This pathway willnot be evaluated further.
The Toxicological Evaluation portion of this section will discuss the possible health hazard fromexposure to the contaminants of concern in surface soil, surface water, and ambient air. Community health concerns will be addressed in the Community Health Concerns Evaluationsection and health outcome data in the Health Outcome Data Evaluation section.
In this section, we will look at the public health implications of the completed exposurepathways (i.e., human exposure occurred or is occurring) identified in the Pathways Analysessection on pages 19 - 20. The Technicon, industrial facility, Owens-Illinois, and CiudadCristiana surface soil, surface water, and the ambient air exposure pathways are consideredcompleted. They, and the contaminants of concern for each pathway, are listed in Table 8 onpage 62.
Typically, the toxicological evaluation in a public health assessment is a comparison of theexposure dose (i.e., the amount of a substance individuals in an exposure pathway are exposed todaily) to an appropriate health guideline. Usually this is ATSDR's Minimal Risk Levels (MRLs)or EPA's Reference Doses (Rfd). The MRLs and Rfds are estimates of daily human exposure toa contaminant below which noncarcinogenic adverse health effects are unlikely to occur (2). That means that any exposure dose below the appropriate MRL or Rfd does not represent ahazard to human health. However, for exposure doses above a MRL or Rfd, there is a wide zoneof uncertainty about whether adverse health effects will occur. Therefore, a review of thetoxicological literature is done to determine whether the specific exposure situation represents ahazard to public health. The results of the comparison of exposure doses to health guidelines,and the methodology for calculating the exposure doses is described in Appendix C on pages 66 -68.
For some chemicals, there is uncertainty about which health guideline (MRL or Rfd) is the mostappropriate to use. This is the case for mercury where there are two MRLs and a Rfd that couldbe used for ingestion. The reasons why the Rfd for methyl mercury was chosen will bediscussed.
The adult exposure doses for the maximum arsenic levels in the industrial facility and totalmercury levels in the Technicon exposure pathways exceeded their health guidelines. Theexposure doses for pica children (children who eat dirt) exceeded the health guidelines for themaximum arsenic, barium, mercury (at only 1 of 164 sampling locations) and manganese levelsin the Ciudad Cristiana surface soil exposure pathway. The possible health consequences ofexposure to those chemicals which exceeded the health guidelines will be discussed.
For the ambient air exposure pathway, the air contaminant concentrations were compared directlyto inhalation minimal risk levels (MRLs) without any additional calculation. Mercury in both theon- and off-site, and methylene chloride in the on-site ambient air exposure pathways exceededhealth guidelines. Air levels of acetone did not exceed health guidelines, but it will be evaluatedfurther because acetone interacts with methylene chloride. The results of those comparisons willbe discussed.
For the surface water exposure pathway to be of health significance, individuals would have hadto drink about 20 milliliters a day and have dermal contact with water from this portion ofFrontera Creek for 20 years. Because there is no indication of this occurring, this pathway willnot be evaluated further.
The risk of carcinogenic health effects was also evaluated. The limitations and methodology forthe carcinogenic evaluation are described in Appendix C. Results of those calculations indicatethat there is no apparent increased risk of cancer for the three carcinogens evaluated - arsenic,beryllium, and methylene chloride.
Neither the exposure dose nor the cancer risk for beryllium calculated using the maximum levelof this chemical exceeded health guidelines. Thus, exposure to beryllium will not be evaluatedfurther.
The Possibility of Health Consequences
The possible health consequences of those exposure doses that exceeded healthguidelines--arsenic, barium, total mercury, methylene chloride, and manganese are described inthe following paragraphs. The risk of cancer from exposure to methylene chloride in the airon-site will also be discussed.
Health assessors determine the possibility of health consequences by comparing the exposure tothe results of epidemiologic evaluations of human exposures to a chemical. If valid human dataare not available then information from properly conducted animal studies are used. The type ofdata used for an evaluation is indicated for each chemical.
Following ATSDR guidance, the possibility of health consequences from exposure to arsenic,barium, and manganese will be evaluated even though all the concentrations of barium andmanganese, and all but two of arsenic, are considered background or naturally occurring.
As described in the Pathways Analyses section, there are two separate exposure situations forarsenic, one involving soil at six industrial facilities and the other, soil at Ciudad Cristiana. Eachsituation will be discussed separately. Also, as was discussed in the Pathways Analyses section,the arsenic levels for all but two of the samples are within what is considered background ornatural. Therefore, only the samples above natural levels (one each at the Coloron and PeerlessTube facilities) can be considered actual contamination. For both exposure situations, there wasno increased risk of cancer from exposure to arsenic.
Industrial Surface Soil
Based on human data, there appears to be no possibility for noncarcinogenic health effects foradults exposed to arsenic at the six industrial facilities where arsenic levels were abovecomparison values (20). The exposure dose for the maximum concentration at the six facilities(430 ppm), is only slightly above the no observed effects level for a population which drankarsenic-contaminated water for 45 years. The exposure dose is about 20 times lower than thelevel where the first health effects (slight effects on the skin) were observed in the samepopulation. Exposure doses for children and pica children are not evaluated here, because, asdescribed in the Environmental Contamination and Other Hazards section, the possibility ofexposure was nonexistent.
Ciudad Cristiana Surface Soil
Based on human data, there appears to have been little possibility for noncarcinogenic healtheffects for children who regularly ate large amounts of soil (pica) at Ciudad Cristiana (20). Theexposure dose for the maximum concentration of arsenic in Ciudad Cristiana soil (4.1 ppm) isabove the no observed health effects level for a population which drank arsenic-contaminatedwater for 45 years. However, the exposure dose is about 6 times lower than the level where thefirst health effects (slight effects on the skin) were observed in the same population. This,combined with the shorter exposure period (6½ years vs. 45 years, indicates that it is unlikely forhealth effects to occur. There is no possibility for health effects for adults and other childrenbecause their exposure doses were much smaller.
As described in the Pathways Analyses section, there was one exposure situation for barium andit involved soil at Ciudad Cristiana. As was discussed in the Pathways section, the barium levelswere within what is considered background or natural.
Based on animal data, there appears to be some possibility that noncarcinogenic health effectscould have occurred for children who regularly ate large amounts of soil (pica) at CiudadCristiana (21). The exposure dose for pica children for the maximum concentration of barium inCiudad Cristiana soil (405 ppm) is above the no effects level and is 3 times greater than thelowest effects level observed in animal studies. This observed lowest effect was an increase inblood pressure. There is some possibility that health effects could have occurred in smallchildren consuming large amounts of soil, but it is difficult to predict this because of theuncertainties involved in using animal data to predict human health effects. There is nopossibility for health effects for adults and other children because their exposure doses weremuch smaller.
Based on human data, there appears to be no possibility that noncarcinogenic health effects couldhave occurred for children who regularly ate large amounts of soil (pica) at Ciudad Cristiana(22). This conclusion is based on an assumption that manganese ingested from soil would besimilar in toxicity to manganese in food. There are data that indicate that the toxicity ofmanganese in food is much less toxic than manganese in water.
Manganese is considered an essential nutrient with a minimum intake of 2 - 3 milligrams/day(mg/d) (22). Extensive dietary investigations indicate that average daily intake varies from 2.0 -8.8 milligrams/day. Those data were used to derive the no observed adverse health effects levelsused as a health guideline. The dietary investigations did not identify an adverse effect inhumans. Animals exposed to manganese in their food were not affected until doses exceeded800 mg/kg/day.
However, in a study done in Greece, mild neurological effects were observed in elderlyindividuals exposed to 0.06 mg/kg/day of manganese in drinking water after many years ofexposure (22). The no observed adverse health effects level for the same population was 0.005mg/kg/day. Children and younger adults exposed to the same levels as elderly individuals werenot affected. Several studies indicate that manganese in water is much more readily absorbedthan manganese in food.
Based on this documented difference, it is assumed that manganese ingested in soil would be nomore toxic than manganese in food. From this assumption, it follows that the levels observed inCiudad Cristiana soil would not cause any adverse health effects.
As described in the Pathways Analyses section, there are four separate exposure situations formercury plus one combined exposure. Two exposure situations involve soil (one at Techniconand the other, soil at Ciudad Cristiana) and the other two involve on- and off-site ambient air. There is a combined exposure situation where workers at Technicon were likely exposed tomercury in soil and in the air. Each situation will be discussed separately.
Choice of Health Guidelines
In this public health assessment, the health guideline for methyl mercury will be used foringestion because it is based on human exposures.
Technicon Surface Soil
The past exposures to mercury in the soil at the Technicon facility alone, likely did not result inadverse health effects. This conclusion is based on the lack of exposure by children, and the lackof a plausible exposure situation for adults. For children to have been at risk of harm, they wouldhave had to have been on the facility all day every day. The Technicon facility is fenced andguarded so this is extremely unlikely. For adults to have been at risk, they would have had tohave contact with the soil around the raw materials storage area all day every work day. Thenumber of individuals who meet these conditions is probably small. Combined air and soilexposures to mercury are discussed briefly below.
Ciudad Cristiana Surface Soil
It is very unlikely that past exposures to mercury in Ciudad Cristiana soil resulted in any adversehealth effects. This is based on the lack of a reasonable exposure situation. Mercury levelsabove the health guideline were found at only 1 of 164 sampling locations (4). Two othersampling locations near this one, were well below the health guideline. This indicates that atCiudad Cristiana exposure to mercury above the health guidelines occurred only at this onelocation. Even at this one location, only a child who ate 5 grams or more of soil a day from thisplace would be at risk of harm. This appears unlikely.
In addition to the lack of significant exposure, the mercury at this one location is about 86%inorganic mercury (717 ppb inorganic/836 ppb total mercury) (4). The rate of absorption foringested inorganic mercury is much lower than for ingested organic mercury (16). Thus,inorganic mercury is less toxic per unit concentration than organic mercury.
On-Site Ambient Air
The maximum (2.2 µg/m3) and average (0.4 µg/m3) concentrations of mercury in the on-site(industrial area) air could have represented a health hazard if those levels are an accuratemeasurement of typical concentrations and if exposure continued regularly for a year or more. However, as discussed on page 12 in the Environmental Contamination and Other Hazardssection, the only available analytical method is not very accurate. In addition, air samples wereprobably not taken on enough days and over a sufficient portion of the day to accurately identifytypical exposure levels. Based on data from long-term worker studies, chronic exposure (a yearor more) to mercury at the levels observed on-site could result in slight muscle tremors (16).
Off-Site Ambient Air
It is not possible to determine whether the maximum (0.09 µg/m3) and average (0.07 µg/m3)concentrations of mercury measured in the off-site air represent a health hazard. As discussed on page 12, the only available analytical method is not very accurate. In addition, air sampleswere probably not taken on enough days and over a sufficient portion of the day to accuratelyidentify typical exposure levels at Ciudad Cristiana.
There was probably a health hazard from exposure to mercury for individuals at the Techniconfacility through ingesting contaminated soil and breathing contaminated air. As describedpreviously, there is a great deal of uncertainty about air exposure levels.
Methylene Chloride and Acetone
It is possible that there could have been health consequences from exposure to methylenechloride in the Frontera Creek area, especially at the Squibb facility from 1983 - 1985. However,this conclusion is uncertain because of difficulties in identifying methylene chloride exposurelevels for the Frontera Creek area and in evaluating the toxicity of methylene chloride. Asdiscussed on page 21, exposure levels for 1983 - 1985 were estimated from 1992 monitoringresults and TRI data. The toxicity of the measured levels is hard to evaluate because ofdifficulties in interpreting data from animal and human investigations, and in evaluating theinteraction between acetone and methylene chloride.
Data from human and animal investigations on the toxicity of methylene chloride is confusingand contradictory. An example of this is the no adverse observed effect level for acute workerexposures (i.e., less than 15 days) of 100 ppm (353,000 µg/m3) which is lower than the 475ppm (1,676,750 µg/m3) level seen for chronic exposures (i.e., greater 365 days). This is theopposite of what is usually observed. In addition, the animal and human studies are of onlyadequate quality which increases the uncertainty about whether the studies identify a "safe"exposure level. This increased uncertainty results in greater margins of safety being used toderive the acute (1390 µg/m3 ) and intermediate (105 µg/m3 ) air environmental media exposureguides (EMEGs). There is, therefore, a large zone of uncertainty between the air EMEG and thelowest level where adverse health effects are first observed.
Another factor in evaluating the possible health consequences of exposure to methylene chlorideis whether the exposure was continuous or intermittent. The lowest observed effects levelsidentified for methylene chloride is the 25 ppm for intermediate exposures (15 - 365 days) forrodents. These results came from a study where the animals were exposed continuously for 100days (23). The higher levels seen in other human and animal studies come from investigationswhere exposure was intermittent (6 - 8 hours a day). Methylene chloride is cleared from the bodywithin 48 hours so a break in exposure allows the body to partly or completely clean the bodyout. Therefore, there is a greater risk of health effects for individuals exposed continuously to thesame level as people exposed intermittently.
There is one animal study indicating that acetone increases the toxicity of methylene chloride(24). The combined exposure to acetone and methylene chloride is probably more toxic thanexposure to methylene chloride alone, even though the acetone concentration is not toxic byitself.
The individuals at risk of having noncarcinogenic health effects related to methylene chlorideexposure, were those who worked at or near Squibb and also lived at Ciudad Cristiana during theperiod of 1983 - 1985. Those individuals may have had continuous exposure to methylenechloride for varying periods and air levels. There also may have been periods when there was noexposure. Given the uncertainty about when health effects would begin to be observed, slightsubclinical (i.e., not harmful) changes in the liver and kidney might have occurred (23). Mild,reversible decreases in vision and hearing might also be seen. However, health effects mighthave occurred only at the highest day time levels and if exposure to lower levels continued athome. The available air monitoring data indicate that this continuous exposure at work andhome occurred only occasionally.
Three other groups of individuals in the Frontera Creek area are, or were exposed to methylenechloride and acetone but probably did not experience any health effects. The first group, peoplewho spent nearly all their day at Ciudad Cristiana (housewives, small children, retired people,etc.), were exposed to about 3 µg/m3 during the day and 4 to over 30 µg/m3 in the evening and atnight. These exposure levels are too low for effects to occur.
The second group are workers at or near Squibb who don't live in the Frontera Creek area. Theywere exposed to day time levels from nondetect to nearly 5000 µg/m3 , but nothing in the eveningor at night. The intermittent nature of this exposure probably allowed their bodies to clean outenough methylene chloride each day so that no health effects were experienced. Incidentally, thisworker exposure is still occurring based on the 1993-94 monitoring. However, Squibb hasreduced methylene chloride emissions and is working with EPA to further reduce them, socurrent exposure levels are lower than previously (10,17).
The third group of exposed individuals were residents of Ciudad Cristiana who left the FronteraCreek area during the day. They were exposed to 4 to over 30 µg/m3 of methylene chlorideduring the evening and night, but were not exposed during the day. Exposure to these relativelylow levels plus the intermittent nature of exposure indicates that no health effects would haveoccurred.
Up to this point in the public health assessment, the focus has been on whether the knownexposures to contaminants in soil or air could result in adverse health effects. Using appropriateavailable data, we will now evaluate whether the people around the site have greater rates ofdisease or higher concentrations of contaminants in their bodies. The discovery of greateramounts of disease or higher levels of chemicals does not prove an association with the site, butdoes indicate the need for further evaluation.
In this section, the guidelines that ATSDR follows for evaluating information from healthoutcome databases will be reviewed, and an explanation of why no cancer or birth defect datawere considered applicable for this public health assessment. Health outcome data of two typeswill be reviewed and evaluated. One will be self-reported health outcomes identified duringpublic availability sessions or from surveys provided by the former residents of Ciudad Cristiana. The other will be blood mercury data for former Ciudad Cristiana residents.
Information from Health Outcome Databases
In a public health assessment, available health outcome databases are identified for the area nearthe site. From those data, ATSDR selects health outcomes for further evaluation based onbiological plausibility or community health concerns (2).
For biological plausibility, the decision to evaluate health outcome data depends on whether acompleted exposure pathway exists for a chemical suspected of causing the health outcome ofconcern (2). The selection of a noncarcinogenic health outcome is based on a review of thetoxicologic literature for that contaminant of concern.
When a contaminant of concern has been identified as a carcinogen, health outcomes for themajor anatomical sites are usually selected for evaluation (2). Designating a chemical as acarcinogen (for purposes of health outcome data evaluation) is based on classification of it as acarcinogen by the National Toxicology Program (NTP), International Agency for Research onCancer (IARC), United States Environmental Protection Agency (EPA), or the United StatesOccupational Safety and Health Agency (OSHA).
A latency period of at least 10 years between exposure and diagnosis has been observed in moststudies of human cancer (25). If exposure began less than 10 years prior to the latest dataavailable, analysis of health outcome data for cancer incidence or mortality is not likely to beuseful, particularly if the exposure level is low.
Even when health outcomes do not meet ATSDR's guidelines for biological plausibility, healthoutcome data can be evaluated to address community health concerns.
An important factor in requesting health outcome data in any situation is the difference in sizebetween the population at risk of exposure to site contaminants and the smallest population unitfor which health outcome data are available (2). For example, adverse health effects due to a sitewould likely not be observed if the population at risk is 100 and the population unit for whichhealth outcome data are available is 100,000.
This last is the reason why cancer incidence data are not described in this public healthassessment. The only available cancer data were for Humacao with a population of 50,000,while the population of Ciudad Cristiana was around 1,500 (4).
If there had been cancer data more applicable to the former Ciudad Cristiana residents, it wouldhave been evaluated. Cancer is a biologically plausible health outcome because there was aprobable human carcinogen, methylene chloride, in a completed exposure pathway.
Developmental effects are considered biologically plausible health outcomes. The availableblood mercury data (see page 34) indicate that some women were exposed to mercury at highenough levels to cause developmental effects. As discussed earlier, mercury is not a site-relatedcontaminant. However, the developmental effects expected (delayed motor development anddelayed language development) from the observed blood levels are not recorded on any availabledatabases, making any evaluation impossible.
Self-Reported Health Outcomes
Data on self-reported health outcomes were obtained from individuals who came to publicavailability sessions and from surveys of the former residents of Ciudad Cristiana. A few of thehealth outcomes reported at the public availability sessions were confirmed by review of medicalrecords provided by the individual. Otherwise, the reported health outcomes were not confirmed. Self-reported health outcomes, obtained by voluntary participation, indicate which diseases orsymptoms may be occurring among a population, but not the frequency.
Data from Public Availability Sessions
At the public availability sessions held in the Vista Hermosa subdivision of Humacao on January13 and 15, 126 ex-residents were interviewed and information was obtained on perceivedsite-related health problems for them and their families (a total of 448 individuals) (10). Someindividuals brought medical records which were reviewed by ATSDR staff.
Health problems reported by the ex-residents by their relative occurrence (most to least reported)were (10): rashes; bleeding gums and loose teeth; headaches; joint and back pain; asthma;mental confusion/memory loss; nervousness; allergies/immune problems; other problems; hairloss; dizziness; fatigue; respiratory illness; poor vision; hypertension; tremors; cardiovasculardisease; reproductive problems; diabetes; numbness; lupus; and cancer.(1) Most of those interviewed identified the probable source of their health problems as mercuryexposure while living at Ciudad Cristiana.
Data from Information Forms
In 1987, the Ex-Ciudad Cristiana Residents' group collected a variety of information from theparticipants in the group's lawsuit against the developers of the subdivision. Self-reported healthoutcomes were included in this information. The former residents group provided ATSDR withcopies of the information forms for 233 households. There were summary data for eachhousehold and for each member of the household. ATSDR selected for review by aspanish-speaking physician the forms for 72 of the 233 households. From the individual forms,the relative occurrence of the self-reported health problems was identified. Also included on theindividual forms were self-reported results of blood mercury testing. The household informationforms also contained non-health concerns.
Data on 259 individuals were identified. The health problems reported in 1987 were similar tothose mentioned at the public availability sessions in 1993 except for some differences in relativeoccurrence. For example, the top five reported illnesses/symptoms were headaches, joint andback pain, other problems (including learning disabilities, infections, and hearing problems),bleeding gums and loose teeth, and rashes. At the public availability sessions, the top five were rashes; bleeding gums and loose teeth; headaches; joint and back pain; and asthma. There wereno reports of diabetes and lupus in 1987, but there were a few in 1993. Abdominal problemswere reported in 1987, but not in 1993.
The results of blood mercury testing for 127 residents were reported on the individuals forms. The mean was 1.87 µg/dl with a range of 0 to 11.7 µg/dl. The mean reported here is lower than3.9 µg/dl identified in the evaluation of blood mercury data that follows this section.
A major non-health concern was a problem with the way the homes were constructed which wasreported by 76.4% of the households (55/72). Another problem was a concern about insectswhich was reported by 47.2% of the households (34/72).
The years of residency at Ciudad Cristiana was also included on the household forms. Theaverage length of stay was 4.3 years.
Evaluation of Self-Reported Health Outcomes
The initial focus of this evaluation will be on the possible relationship of the reported healthproblems with exposure to mercury. Based on a review of the literature, the following symptomshave not been identified as being associated with mercury exposure: joint and back pain, asthma,cardiovascular disease, diabetes, lupus, and cancer (16,26). The other symptoms have beenassociated with exposure to elemental mercury vapor, or organic or inorganic mercury, but alsoto numerous other chemical and biological agents.
It is very unlikely that any of the symptoms reported to be occurring when the interviews wereconducted in January 1993 are due to mercury exposures that occurred while living at CiudadCristiana. While mercury can cause severe lifelong effects if exposure levels are high enough,the reported symptoms would have disappeared in the eight years since the possible exposure atCiudad Cristiana ended (16). More importantly, as discussed on page 15, the mercury levels atCiudad Cristiana were typical for Puerto Rico.
When a physician attempts to determine whether a individual's symptoms are due to mercury,good evidence of exposure is required (26). This evidence can be indirect such as working withmercury or eating mercury-contaminated fish, or direct such as blood, hair, or urineconcentrations at levels of health concern. None of the 290 Ciudad Cristiana residents, whoseblood mercury tests were evaluated in this public health assessment, had such levels. However,as described in the next section, some women had blood mercury levels which could haveresulted in delayed motor and language development in their children exposed in the womb.
If not mercury, what could be causing the symptoms and illnesses reported? The concentrationof methylene chloride in the air may have been high enough to cause some illness. When theresidents first began to complain of health problems in 1984, air pollution was mentioned as apossible cause (11). Interestingly, mosquitoes were also identified as a problem. Some of thereported health problems are also consistent with mosquito-borne or other infectious diseases(27). Another possible cause for symptoms such as headaches and dizziness is the stress. Manyformer Ciudad Cristiana residents were concerned because they did not own the homes they wereliving in at Vista Hermosa and might have to move at any time (10). Stress from concern aboutsuch issues has been related to illness (28).
Data from Analyses of Blood for Mercury
Testing blood, tissue, or other body fluids from individuals living around a site for possible sitecontaminants can identify whether exposure has occurred and if contaminant concentrations arehigh enough to cause adverse health effects. Many of the over 1000 residents of CiudadCristiana were tested for mercury, with the vast majority of the analyses being for mercury inblood. Urine and hair were tested for mercury; blood from a few individuals was also tested forlead.
Of those various biological analyses, ATSDR has in its records the individual results for severalhundred blood mercury analyses. This data set was the only one suitable for evaluation in thispublic health assessment because there was adequate information on quality control/qualityassurance and a fairly good number of results available. An in-depth evaluation of those bloodmercury data and three tables listing results can be found in Appendix D (pages 69 - 79). Asummary of that evaluation follows.
ATSDR's review of blood mercury data for about 290 Ciudad Cristiana residents indicates thatthe average blood mercury level was 3.9 micrograms of mercury per deciliter of blood (µg/dl). The most likely source for this mercury was local fish. The mercury levels in Frontera Creekarea fish are typical for Puerto Rico. However, Frontera Creek area fishermen eat large amountsof fish a day which could result in mercury exposure. As described on page 74 in Appendix D,other sources of mercury exposure are possible.
Individual blood mercury concentrations were not above the 20 µg/dl level where health effectsappear to begin for this type of mercury exposure (29,30). However, health effects might haveoccurred in some of the unborn babies of mothers whose blood mercury levels were above 4.0µg/dl. Twenty-seven of the 64 women (42%) of child-bearing age tested had levels above 4.0µg/dl. The health effects that might be observed at mother's blood mercury levels of 4 - 20 µg/dlare delayed motor development and delayed language development (29-33). This means thatchildren exposed to these levels while in their mother's womb may have walked later than thenormal maximum of 19 months, or talked later than the normal maximum of 26 months (33). Astudy done of Iraqi children indicates that about 25% of the children with maternal bloodmercury levels of 4 - 20 µg/dl would have walked or talked later; the rest with this range ofmaternal blood levels had no adverse health effects. It is not known whether thosedevelopmental delays would be permanent or temporary. At maternal levels of 4 - 20 µg/dl,more severe effects such as mental retardation, deafness, blindness, microcephaly (small brain),and cerebral palsy are not observed (30,31).
The evidence available to ATSDR indicates that only those of Ciudad Cristiana residents whohad their blood mercury analyzed privately (i.e., the set of results evaluated in this document)appear to have been at risk in January - June 1985 of having children who had been exposed togreater than 4.0 µg/dl of mercury in the womb. Only 6 of 993 Ciudad Cristiana residents testedby the Puerto Rico Department of Health had blood mercury levels above 5 µg/dl, and thusprobably few, if any, women would have levels above 4.0 µg/dl. Because both sets of bloodmercury testing were conducted in the first half of 1985, conclusions about the number ofchildren at risk of developmental effects can only be made for that time period.
A major limitation with the blood mercury data reviewed is that the participants wereself-selected which could result in a weighing of the results either to the high or low side. Thedifference of the percentage of results above 5 µg/dl between the Puerto Rico Department ofHealth (PRDH) study (0.6% - 6/993) and this set of results (29.0% - 84/290) indicates that theremay be such a problem.
C. Community Health Concerns Evaluation
The community health concerns identified during the public meeting are addressed in thissection. Those concerns obtained at the public availability sessions were evaluated in the HealthOutcome Data Evaluation section (pages 32 - 33).
CONCERN - ATSDR's conclusion that the mercury levels in the Frontera Creek area,except for the Technicon facility, are typical for the southeast region of Puerto Rico is notcorrect. The three main lines of argument against ATSDR's conclusion were:
- Most of the mercury in the soil and sediment while people lived in Ciudad Cristiana (1979 -1985) may have evaporated or migrated off-site through surface water runoff by the time thesampling for the RI was done.
- There was good evidence that Ciudad Cristiana was grossly contaminated. Evidence for thisincludes statements by the Puerto Rico Environmental Quality Board (PREQB) in 1979 thatFrontera Creek was contaminated, the reported use of creek sediments as fill for CiudadCristiana, and the reports of illness among the workers who moved the sediment.
- The elevated blood mercury levels in Ciudad Cristiana residents in 1985 prove that they wereexposed and that Ciudad Cristiana was contaminated.
ATSDR's response to those three lines of argument are as follows.
1) It is very unlikely that most of the mercury in the soil and sediment in the Frontera Creek areawould have evaporated or migrated off-site through surface water runoff between the time whenthe residents left in 1985 and the time the RI sampling was done in January and February 1988. This conclusion is based on the sampling data from 1978-1985, and the biochemical cycling ofmercury in the environment (4,16).
The volatility (ability to evaporate) and solubility (ability to dissolve in water) varies greatlydepending on the type of mercury (16,29,34). Elemental mercury and a few forms of organicmercury evaporate easily, but the most common forms of inorganic and organic mercuryevaporate very slowly. Elemental mercury and the most common forms of inorganic mercurydissolve in water very slowly, while organic mercury dissolves easily. Therefore, the rate whichmercury leaves an area through evaporation or migration in surface water runoff would dependon the source of the mercury in the soil and sediment. For the Frontera Creek area, the sources were the Technicon facility and naturally-occurring mercury (4).
As indicated in Table 16 (Appendix E), the soil mercury levels in 1985 identified by EPA aresimilar to those in 1988. As discussed, the soil mercury levels at Ciudad Cristiana are withinbackground (natural) ranges for the area, and not due to contamination (4). The levels ofnaturally-occurring mercury would remain relatively constant through time due to thebiochemical cycling of mercury (16).
Technicon used several inorganic mercury compounds including mercuric acetate and mercuricchloride and one organic compound, thimerosal, which is a ethyl mercury derivative in theiroperations (4,35). Until 1979, wastes from those compounds were discharged into FronteraCreek via the Technicon ditch.
All of the compounds discharged into Frontera Creek have slow evaporation rates; all but theethyl mercury compound (thimerosal) dissolve very slowly in water (16,29,34). The amount ofthimerosal used at Technicon was small (about 125 pounds/year) compared to the inorganiccompounds (4). Therefore, because the compounds from Technicon do not evaporate or dissolvequickly (except for thimerosal), there should be relatively little change between the levelsdetected in Frontera Creek sediments in 1978 and 1988.
Mercury levels in sediment are similar in 7 of 9 sampling events including the remedialinvestigation (Table 17 and 18, Appendix E). The October 1979 sampling by EPA identified significantly higher mercury levels. However, it is likely that those higher levels in October 1979are due to the disturbance of Frontera Creek sediments by dredging in February 1979 (36). Thiswould have uncovered sediments contaminated before the discharging of mercury compoundswas discovered in 1978. All discharges stopped in 1979. Mercury levels were much lower insamples taken in February 1979 and in 1983, when new sedimentation would have covered,diluted, or mixed with older sediments. The results of the other sampling event with some highermercury levels, conducted in March 1985 by EQB, are questionable because of the same qualitycontrol concerns as the EQB's Ciudad Cristiana soil samples obtained at the same time (1).
2) ATSDR does not agree that the evidence presented proves that there was gross contaminationof Ciudad Cristiana by mercury. First, while ATSDR agrees with EQB and EPA that portions ofFrontera Creek were contaminated with mercury, this in no way proves that Ciudad Cristiana wascontaminated. As discussed in the Environmental Contamination and Other Hazards sections,the sampling of Ciudad Cristiana in 1988 was very systematic and thorough. If there had beengross contamination, at least some of the surface or subsurface samples should have shownmercury levels above background; they did not. Also, the subsurface sampling found noindication that sediments had been used as fill (4). In addition, the 1988 RI did not identifyexcess mercury levels in biota. Because mercury bioaccumulates, there should have been greateramounts in the biota in 1988 if there had been gross contamination of the Frontera Creek area in1985.
ATSDR did identify a report indicating that attempts were made in 1979 to deepen FronteraCreek by removing sediment (36). Actions by the Corps of Engineers and EPA stopped thisactivity by the developers of Ciudad Cristiana. There appears to have been no widespread use ofthis sediment at Ciudad Cristiana as indicated by the evaluation of well boring cores which didnot identify any dredge spoils (4).
However, a former Ciudad Cristiana resident told an ATSDR staffer in January 1993 that someof the rich organic material dredged from Frontera Creek was used as top soil in residentialyards.(2) This same individual also related that one of the dredge workers sold several dump truckloads of this rich organic material. If this material was used as top soil, there is no indication thatit was highly contaminated based on EPA's 1985 and 1988 investigations (4). In addition, theconcentration of the mercury in the material used as top soil in 1979 would not have decreasedsubstantially in 1988 because mercury adheres tightly to organic materials (29).
ATSDR was also unable to identify any information confirming reports that the workersremoving the sediments became sick.
3) The presence of mercury in residents' blood in no way proves that Ciudad Cristiana soil wascontaminated. The soil would have to be contaminated above background levels before soilcould be considered a significant source of exposure. The available evidence identifies mercuryin fish as a major, though probably not the only source. As extensively discussed in the HealthOutcome Data Evaluation section of this public health assessment, the average blood mercurylevel of 3.9 µg/dl for Ciudad Cristiana residents is typical for a population with a high rate of fishconsumption. Both the blood mercury results evaluated in this document and the PRDH studyprove that the residents of Ciudad Cristiana were exposed to mercury, but this could be said forjust about any population tested anywhere in the world.
CONCERN - Why is ATSDR planning to use the blood mercury data in the public healthassessment, whereas these data were considered unacceptable by ATSDR in 1986 and 1988Health Consultations?
In the earlier evaluations, ATSDR was unable to interpret the blood mercury data due to a lack ofinformation on collection procedures, quality assurance/quality control (QA/QC), backgroundlevels, and other sources of exposure. We now have information on collection procedures, otherpossible sources of exposure, and background levels. The blood mercury data is evaluated in thepublic health assessment and conclusions were reached, even though the lack of data on QA/QCintroduces some uncertainty.
CONCERN - What, if any, are the health consequences associated with the blood mercurylevels identified in former Ciudad Cristiana residents?
ATSDR's review of blood mercury data for about 300 Ciudad Cristiana residents indicates thatthe average blood mercury level was 3.9 micrograms of mercury per deciliter of blood (µg/dl). The most likely source for this mercury was local fish. The mercury levels in Frontera Creekarea fish are typical for Puerto Rico. However, Frontera Creek area fishermen eat large amountsof fish a day which could result in mercury exposure.
Individual blood mercury concentrations were not above the 20 µg/dl level where health effectsappear to begin for this type of mercury exposure. However, health effects might have occurredin some of the unborn babies of mothers whose blood mercury levels were above 4.0 µg/dl. Twenty-seven of the 64 women (42%) of child-bearing age tested had levels above 4.0 µg/dl. The health effects that might be observed at mother's blood mercury levels of 4 - 20 µg/dl aredelayed motor development and delayed language development (29-33). This means thatchildren exposed to these levels while in their mother's womb may have walked later than thenormal maximum of 19 months, or talked later than the normal maximum of 26 months (33). Astudy done of Iraqi children indicates that about 25% of the children with maternal bloodmercury levels of 4 - 20 µg/dl would have walked or talked later; the rest with this range ofmaternal blood levels had no adverse health effects. It is not known whether thosedevelopmental delays would be permanent or temporary. At maternal levels of 4 - 20 µg/dl,more severe effects such as mental retardation, deafness, blindness, microcephaly (small brain),and cerebral palsy are not observed (30,31).
CONCERN - What, if any, are the health consequences posed by past and present airemissions from one of the industrial facilities at Frontera Creek?
It is possible that there could have been health consequences from exposure to methylenechloride in the Frontera Creek area, especially at the Squibb facility from 1983 - 1985. Theindividuals at risk of having noncarcinogenic health effects related to methylene chlorideexposure, were those who worked at or near Squibb and also lived at Ciudad Cristiana during theperiod of 1983 - 1985. Slight subclinical (i.e., not harmful) changes in the liver and kidney mighthave occurred in these individuals. Mild, reversible decreases in vision and hearing might alsohave been seen. However, this conclusion about methylene chloride is uncertain because ofdifficulties in identifying methylene chloride exposure levels for the Frontera Creek area and inevaluating the toxicity of methylene chloride.
Three other groups of individuals in the Frontera Creek area are, or were exposed to methylenechloride and acetone but probably did not experience any health effects. The first group werepeople who spent nearly all their day at Ciudad Cristiana (housewives, small children, retiredpeople, etc.), the second were workers who didn't live in the Frontera Creek area, and the thirdwere Ciudad Cristiana residents who worked outside of the Frontera Creek area. Exposure tosome workers is still occurring. Squibb has reduced methylene chloride emissions and isworking with EPA to further reduce them.
For more information on methylene chloride emissions see page 22 of the Pathways Analysessection and for a discussion of methylene chloride toxicity see pages 29 - 30 of the ToxicologicalEvaluation section.
1.After the results above were compiled, one of the Spanish-speaking ATSDR staff related thatthe Spanish word for fatigue is often used as a synonym for asthma (10). Therefore, the numberwho believe they have asthma is actually higher than that indicated.
2.This information was given to Mr. Antonio Quiñones during ATSDR's January 1993 sitevisit.